Bioatomspace agent

[besSveta]

bio-ai-agent uses llm to convert query in natural language to query in metta run bio-ai.metta from examples folder ( the working directory is examples)

[Necr0x0Der]

I'm not sure it is a good place for bio-ai agent.

• The example will not work without downloading bio-atomspace, so it is mostly useless as a general example for metta-motto. In particular, test_bio_ai_agent.metta contains "/media/sveta/hdisk4/singnet/bioatomspace/bio_atomspace_subset/valid"
• Is there a real need for bio_ai_agent in Python? Does it contain any fundamental to be included into metta-motto library itself? I guess it automates loading all bio-atomspace files. It's ok doing this in Python, although I'd propose to have a version with explicitly loading all the files using load-ascii in msa-file, or to have one extension Python with the function like "load-biospaces", which will automate loading of all files. Or is the bio_ai_agent doing something in addition? In general, I'd put all the files in the separate folder. It can be in metta-motto (possibly, a subfoder in examples), and make it self-contained (add small subsets of bio-atomspace files sufficient for queries).

[Necr0x0Der]

Minor proposals:

• move bio_ai_agent to examples/bio_ai
• rename help_functions.py to utils.py or string_utils.py if there are no plans to add other type of helper functions there